Precision Landing and elevated take off point

[Mauri59]

Is there any relation, between the effective ground level and the height of a small (1sq mt), elevated take off point?
Until a couple of days ago, the popup banner text, that one with the little square to mark, warned that a minumum "altitude" of 6 meters was required, to store the visual image, in order to obtain a precision landing in RTH.
Today, while taking off from such an elevated platform (new experiment), I noticed the text in the banner indicated a substantially reduced height, of only 3.7m.
Wondering now (it puzzles) if such reduced height is caused by my small elevated platform, or instead, if it's something related to new fw v1.03.0800.
Incidentally, the platform is much higher from ground level, to justify 3.7m as the aritmetic difference (platform is abt 4.5m), infact 6 - 4.5 = 1.5
By the way, the GPS calibs have been done on the elevated platform.
Any idea?
(can't check it now, since I can' fly until tomorrow)

 

[AlanTheBeast]

1) Per the manual, including the latest (p. 15), Precision Landing requires that you rise vertically 10 metres over the takeoff/landing point.

2) RTH does not require the image. It requires a GPS position. This is why one should not start the motors until GPS is acquired and the home point recorded. GPS accuracy should always be considered as not better than 10 metres even though it will usually be better than 3m (horizontally). If you are in a cluttered environment, canyons (real or urban) the error can be 10's of metres.

Since the Precision Landing function makes a 3D image with 2 cameras, it's possible that an elevated platform might give it a confusing image. I can't confirm this - but food for thought.

There is no GPS calibration, so not sure what you're on about with that. If you mean you wait until GPS is acquired before taking off, then fine. If you want more accuracy there, wait for more satellites. (As an aside, in case you don't know, the drone does not use GPS altitude as reference but instead it uses its internal baro altimeter).

Curious: Why are you using an elevated takeoff platform?

 

[Mauri59]

Yes, the manual says 10mt, but 6m worked same fine (I did a number of test). Probably that's a mistake, an advice surpassed by newer mavic features. Otherwise what's the meaning of the message in the banner?
I didn't express well, I meant the image used for precision landing during an RTH.
I think a 3D image, should just avoid these uncertainty. What can it be, when taking off in a mountain area, where "flat" areas might be as small as my "platform"?
The ground (my backyard) is temporary unavailable because of works (gardening stuffs on).

 

[AlanTheBeast]

What - precisely - is the message it says at 3.7 metres - or whatever? Is it the "Home Point Recorded" message?

There is no difference in the Precision Landing image for RTH or any other purpose. RTH will work fine if there is no image at all (but will be less accurate - for example at night).

 

[Mauri59]

First, I'm talking about DJI go4 app msg.
The message, states that the drone has to reach the specified height, to insure a precision landing. Marking the square and sliding the virtual switch, the mavic takes off and reaches that height, automatically. (leaving the square unmarked, instead, it takes off and hovers much lower)
I know the RTH is possible also whitout picture, but this is not the point, here.
Just wondering if the (much) lower height, 3.7m to take the shot, that one I've had in the banner'app today, was consequence of the new FW (as probably was previously, with 6m, instead of 10m), or because of the elevated platform.

 

[AlanTheBeast]

I see. I never use automatic take off features (and wonder why people do). I'll try to remember to check it out next time I'm out.

 

[Mauri59]

Update...
After a number of trial and check, the "precision landing" feature of my Mavic seems to be fooled by the height of nearby structures and somehow, it calculates, as ground, an average (wrong) height, among effective ground (abt 4.5mt below the platform), the platform itself (one square meter) and a wodden "tool house" roof, which is a couple of meters away.
I think there's no way to land automatically, if taking off area doesn't extend enough (and flat), around the Mavic.
When the Mavic tries to land, at "its estimated ground", which is lower height than platform, the bottom sensors see the platform as an obstacle and the descend is stopped.
I tried to find out how much it has to extend the "precision landind area" to avoid the system isn't fooled, but I didn't find this information anywhere, both in the specs and browsing the Internet.
Anybody knows or tested himself?

Today, while taking off from such an elevated platform (new experiment), I noticed the text in the banner indicated a substantially reduced height, of only 3.7m.
Wondering now (it puzzles) if such reduced height is caused by my small elevated platform, or instead, if it's something related to new fw v1.03.0800.
Incidentally, the platform is much higher from ground level, to justify 3.7m as the aritmetic difference (platform is abt 4.5m), infact 6 - 4.5 = 1.5
By the way, the GPS calibs have been done on the elevated platform.